Vahe Hakobyan
Vahe Hakobyan
> > @sayakpaul worked yesterday on the tests. Hit a wall. Then tried to run tests for the text_to_image and hit the same wall: > > Had it been the...
yes, read the last paper by Gatys
@Heartsie what parameter set did you use ?
@Heartsie did I get it correct that you left other parameters untouched ? like lambda_style, lambda_feat and so on...
@yanhn you subtract 120 just before you feed it to VGG network. For the style it's subtracted by vgg.preprocess method :)
yes, they are measuring the strength of the style, content and spatial smoothness in the outcome image accordingly.
The paper only says about lambda_tv between 1e-4 and 1e-6....about others I don't know anything
I have the same issue, fighting against it 12 hours already
@jyoung105 found the problem it is actually because of algumentations : specify the version 1.4.3, two days ago they release 1.4.4 and it breaks everything